Learning and exchange rate dynamics: A theoretical model

This paper focuses on the determination of the spot and the forward exchange rate assuming that there is uncertainty about the development of a fundamental. It is argued that, in this case, agents in the foreign exchange market try to infer the true parameter values from newly available observations. It is assumed that agents apply OLS estimations to derive an unknown parameter. When new data are available agents update their estimate and thus go through a learning process. It is shown that the learning process causes a bias of the forward rate due to systematic expectation errors and that it also leads to an increase in exchange rate volatility. Only in the long run and in the absence of uncertainty does the exchange rate coincide with the level implied by rational expectations.